Introduction
Research has identified and attempted to explain the significant correlation between experience of natural disaster and comorbid psychological consequences, including posttraumatic stress disorder (PTSD) and depression.Reference Bell and Folkerth 1 - Reference Felix and Afifi 6 Moreover, the experience of natural disaster often is related to increased rates of intimate partner violence (IPV) among survivors.Reference Norris, Friedman, Watson, Byrne, Diaz and Kaniasty 7 - Reference Larrance, Anastario and Lawry 11 However, knowledge of the specific mental health consequences of disasters experienced by women is limited,Reference Picardo, Burton and Naponick 12 and even less has been done to identify the risk and mitigating factors that may be used to identify and provide relief for women in disaster situations.Reference Corrarino 13 , Reference Taft, Monson, Schumm, Watkins, Panuzio and Resick 14 If research is to be relevant in impacting national policy and legislation for natural disaster preparation and relief, the health consequences faced by vulnerable populations post-disaster must be delineated and data made accessible to policymakers. Therefore, the purpose of this study was to assess the existing literature in order to identify the most prevalent mental and psychological health consequences experienced by women following natural disasters in the United States and to understand the link, if any, between post-disaster trauma and IPV.
Natural Disaster in the United States
Definitions for disaster vary, though the term has been understood as a “calamitous event that effects a large population and generally results in injury, death, and destruction of property.”Reference Rosborough, Chan and Parmar 15 Since 1990, the United States has experienced an average of 53 major disasters each year, with approximately 11 disasters annually warranting federal assistance under the Robert T. Stafford Disaster Relief and Emergency Assistance Act. 16 As the United States continues to experience both population growth and continued development/urbanization, the probability that a given individual will directly be exposed to, or affected by, a natural disaster over the course of their lifespan is likely to increase dramatically; the risk inherent in geographic vulnerability is only exacerbated by climate change. 17
Disaster and Mental Health: A Gendered Approach
Significant headway has been made in delineating the most common human health consequences of natural disasters. Norris et al’s review of empirical disaster research, published in 2002, provided a careful synthesis of 20 years of disaster research literature. Focusing on mental health outcomes, Norris’s article provided significant insight into the disaster experience internationally and at all stages of the life cycle, and included 42 samples of adult survivors experiencing natural disasters in the United States.Reference Felix and Afifi 6 Information and data on the mental, physical, and psychosocial sequelae triggered by natural disasters among many populations have since proliferated,Reference Bell and Folkerth 1 , Reference Greenough, Lappi and Hsu 18 - Reference Boscarino, Hoffman, Adams, Figley and Solhkhah 20 and subsequent research has linked natural disasters to a variety of mental health concerns, including PTSD, depression, anxiety, elevated stress and suicide rates, sleep problems, and drug use.Reference Bell and Folkerth 1 , Reference Adeola 2 , Reference Ginexi, Weihs and Simmens 21 - Reference Timpson, Ratliff and Ross 26
Within this gendered context, women suffer disproportionate rates of IPV under normal circumstances, but when combined with the stressors experienced post-disaster such as disrupted social networks and loss of resources, the potential for IPV can increase.Reference Leon, Hyre, Ompad, Desalvo and Muntner 22 When looking at extant disaster literature, however, a focus on the health consequences of disasters specific to women is rare, and a study comprehensively assessing those health consequences most likely to be experienced by women following natural disasters in the United States is needed. Articles not restricted to (1) health consequences experienced by women (2) following natural disasters (3) in the United States limit researchers’ ability to identify the specific health risks that can inform future policy development and implementation.
Moreover, much of the existing research, and particularly that predating Norris’s 2002 article, continues to use dated research; many of the studies included in Norris’s article were conducted before a gendered approach to disaster research became a significant academic and policy goal. For this reason, the current study reviewed literature published after 2000 in order to allow for some overlap with the Norris article. Scoping reviews serve to map a broad section of the evidence base around a topic, characterize the quality of the evidence base in terms of type of studies, allowing for the identification of gaps in knowledge and formulation of research questions. Reference Wu, Stevens, Tennen, North, Grady and Holzer 27 Therefore, a scoping review was conducted to determine the mental health and safety consequences of natural disaster most prevalent in women living in the United States.
Methods
Search Strategy
A search strategy was developed with the assistance of a health sciences informationist at a large research university in the United States. Inclusion and exclusion criteria were determined a priori. Six databases (PubMed [National Center for Biotechnology Information; Bethesda, Maryland USA]; PsycINFO [American Psychological Association; Washington DC, USA]; Cochrane [The Cochrane Collaboration; Oxford, United Kingdom]; JSTOR [ITHAKA; New York, New York USA]; Web of Science [Thomson Reuters; New York, New York USA]; and databases available through ProQuest [Ann Arbor, Michigan USA], including ProQuest Research Library) were searched for articles published from 2000 through 2015 that assessed the health consequences most likely to be experienced by women following natural disaster. Using ProQuest, which provides advanced indexing, allows for a more diverse number of sources in the area of study to be reached.
The core search was conducted in PubMed and combined keywords in the title and abstract with medical subject headings. Other searches were variations of the core search. The initial searches were conducted from February 23 through March 18, 2015 (Table 1; Figure 1).
Figure 1 PRISMA Diagram. Abbreviation: PTSD: posttraumatic stress disorder.
Table 1 Search Terms and Review Criteria
Study Selection
Initial inclusion criteria were (1) English-language articles (2) published between 2000 and 2015 that (3) explored adult survivors’ health outcomes (4) following natural disasters (5) occurring in the United States. Duplicate records, articles with similar keywords but unrelated topics, and non-research articles were eliminated. While the intention was to identify the impacts of natural disaster on women’s health, articles that assessed health consequences in both male and female victims were retained. Natural disasters were distinguished from mass-violence disasters (including gun-related massacres and terrorist attacks) and technological disasters (including nuclear, industrial, and transportation accidents).Reference Felix and Afifi 6 , Reference Arksey and O’Malley 28 Both in-depth studies of a single disaster event and those comparing multiple disaster events or reviewing the literature related to specific complaints were included.
Data Extraction
Information on author, year, title of study, journal, health outcome measured, disaster type and date, study population and sample size, study methodology, aims of study, summarized results and implications, and validated outcome assessment instruments were collected in an Excel (Microsoft Corp.; Redmond, Washington USA) workbook. At each stage of the review process, one researcher reviewed titles and abstracts for possible inclusion and a second reviewer independently assessed the relevance of the initially-selected titles and abstracts for exclusion. Differences in results pulled into a separate file and reviewers resolved them through discussion until consensus was reached. Because the intent of this scoping review was to assess the most common post-disaster mental health symptoms and IPV impacts experienced by women in the United States, an assessment of the quality of studies reviewed was not conducted.Reference Norris, Friedman and Watson 29 However, study designs and levels of evidence were identified as part of this review in order to identify gaps in research as well as appropriate next steps. The Oxford Centre for Evidence-Based Medicine (University of Oxford; Oxford, England) levels of evidence and evidence grading system was used (Table 2 and Table 3).
Table 2 Comparison of Disasters Included in the Review
a Cost in US $.
Table 3 Studies on the Health Impacts of Natural Disasters in the United States
Abbreviation: IPV, intimate partner violence.
Results
Of the 3,878 articles originally identified, 332 were removed as duplicates. Of the 3,546 remaining, 149 articles met the initial criteria for inclusion. The titles and abstracts of these 149 relevant articles were reviewed. Supporting literature and non-research articles, including reviews, commentaries, and editorials, were labeled and removed (64 articles; 43.0%). Articles focused exclusively on physical health and prenatal/maternal health also were excluded (27 articles; 18.1%). The remaining articles focused on PTSD/mental health impacts and IPV following natural disasters; these were selected for full-text review (58 articles; 38.9%) and constitute the 58 primary literature articles included in this scoping review (Appendix 1; available online only).
The 58 articles under full-text review represented 12 separate natural disasters occurring from 1993 through 2012 (Table 4). The disasters differed significantly in severity, with fatalities ranging from none during the Santa Barbara County (California USA) wildfires in 2008-2009 to 1,833 deaths caused by Hurricane Katrina (Louisiana USA) in 2005.Reference Davis, Grills-Taquechel and Ollendick 5 , Reference Armstrong, Hall, Doyle and Waters 30 Hurricane Katrina was the most-costly natural disaster, with estimated damages of US $108 billion, while the Buffalo Creek, Colorado (USA) wildfire and subsequent flood of 1996 cost only US $20.25 million.Reference Armstrong, Hall, Doyle and Waters 30 - Reference Schrader 32
Table 4 Study Designs Included in Review
a Centre for Evidence-Based Medicine. (2009). Oxford Centre for Evidence-based Medicine – Levels of Evidence. University of Oxford: Oxford, England.
The 58 articles chosen for full-text review were categorized loosely based on identified themes in research topics. These included: (1) PTSD/mental health in non-pregnant women (38 articles; 65.5%); (2) PTSD/mental health in pregnant and postpartum women (10 articles; 17.2%); (3) IPV (six articles; 10.3%); and (4) IPV in conjunction with mental health concerns (four articles; 7.0%; Table 5).
Table 5 Study Characteristics
Abbreviations: GBV, gender-based violence; IPV, intimate partner violence.
The most common design used was post-event, cross-sectional studies (42 articles; 72.4%). Nineteen studies (32.7%) were longitudinal, with 12 articles (20.7%) collecting data both pre- and post-disaster, and seven articles (12.1%) collecting data only after the disaster. Fifty-two articles (89.7%) used data from a single, disaster-exposed group, while three articles (5.2%) used qualitative data. Sample sizes ranged from 40 to 6,309 participants with an average of 552.6 study participants. Follow-up times and points of assessment varied considerably across the studies and ranged from 1-14 days post-disaster to 53-61 months post-disaster (Table 2).
Of the 59 studies, 47 (81.0%) used validated outcome assessment instruments. Several studies drew information from narrative or semi-structured interviews with study participants (13 articles; 22.4%) and surveys or question lists designed specifically for use in that particular study (18 articles; 30.5%).
Discussion
This scoping review exposes significant deficiencies in the extant literature and highlights critical areas of concern. It also integrates evidence on knowledge gaps, where high levels of evidence indicate previous work has had a significant impact on field, while lower levels of evidence suggest a lack of vital research in the area. In particular, although higher rates of post-disaster PTSD, depression, and anxiety for women as compared to men are reported,Reference Bell and Folkerth 1 , Reference Wolf 33 , Reference Brown, Cherry and Marks 34 little progress has been made to standardize research across disaster types and affected populations in order to identify the aggravating and alleviating factors that contribute to mental health concerns, and even less has been accomplished across disciplines to target clinical care towards those most in need, or to influence legislation and disaster relief policy. The dearth of relevant data is most apparent in understanding the compounded relationship between post-disaster psychological illness and increased incidence of IPV and gender-based violence (GBV).Reference Fredman, Monson, Schumm and Adair 10
That women experience significantly higher rates of post-disaster mental illness than men has been well documented.Reference Felix and Afifi 6 , Reference Arksey and O’Malley 28 , Reference Chan, Rhodes and Perez 35 As a gendered approach to disaster research becomes more recognized, however, it increasingly is understood that post-disaster mental illness is particularly acute for women who are also minorities, poor, elderly, or the primary caregivers for their children.Reference Bell and Folkerth 1 , 17 , Reference Galea, Nandi and Vlahov 36 - Reference Kulkarni and Pole 39 These impacts are cumulative with vulnerability to health trauma increasing as socioeconomic status declines.
Adeola’s study of mental health and psychosocial distress in survivors of Hurricane Katrina indicates that African Americans are more likely than their white counterparts to experience an array of psychosocial problems, including feelings of depression, anxiety, and worries about the future.Reference Bell and Folkerth 1 While this study indicated that employment and income are significant inverse predictors of psychosocial distress following natural disaster, minority women are a population least likely to have means of support, such as a steady job or source of income.Reference Bell and Folkerth 1 The sense of powerlessness and helplessness associated with survivors’ perceptions of having been victimized or taken advantage of by opportunists was also a critical factor elevating psychosocial dysfunction.Reference Bell and Folkerth 1 Similarly, among survivors of the 1994 Northridge Earthquake in California, Asian Americans were more than twice as likely to experience psychiatric distress warranting clinic attention than their European American counterparts.Reference Joseph, Matthews and Myers 38 Among pregnant and postpartum women, nontangible loss of resources was a significant predictor of depression, with African American women, women in their mid-twenties, and women without a college degree more likely to report high loss of resources.Reference Rhodes, Chan, Paxson, Rouse, Waters and Fussell 40 , Reference Ehrlich, Harville, Xiong, Buekens, Pridjian and Elkind-Hirsch 41
Several hypotheses for the disparate impact of post-disaster psychological trauma in women have been advanced. Exposure to natural disaster, often measured by survivors’ experience of lacking basic necessities such as food, water, and medical care, have been cited as predictors of poor mental health outcomes following natural disaster.Reference Cepeda, Valdez, Kaplan and Hill 3 , Reference Harville, Xiong, Pridjian, Elkind-Hirsch and Buekens 42 Evacuation, extended displacement, and resource loss, such as the loss of one’s home, also have been correlated to experience of increased symptoms of depression and PTSD.Reference Chan and Rhodes 4 , Reference Wolf 33 , Reference Kim, Plumb, Gredig, Rankin and Taylor 43 However, this has not been clearly established, nor have the reasons why women have consistently higher rates of disaster exposure than men.Reference Bell and Folkerth 1 , Reference Jones, Ribbe, Cunningham and Weddle 44
It has also been posited that survivors’ early symptoms may serve as predictors of poor mental health outcomes.Reference Greenough, Lappi and Hsu 18 , Reference Benight and Harper 19 Because many women assume primary responsibility for housing and childcare in their families, their ability to provide adequate self-care in the immediate aftermath of natural disasters is limited, and post-traumatic stress symptoms may be latent.Reference Cepeda, Valdez, Kaplan and Hill 3 , Reference Galea, Nandi and Vlahov 36 In the immediate aftermath of Hurricane Sandy (2012; New York USA), for example, men and women reported similar rates of post-traumatic stress symptoms, though women were more likely to report recollections of previous disasters and fear of future disaster events.Reference Galea, Nandi and Vlahov 36 Similarly, in a combination wildfire-flood disaster occurring in Colorado, limited gender disparity was found in participants’ acute stress response two weeks after the flood, but significantly greater symptoms of PTSD were reported in female participants one year later.Reference Greenough, Lappi and Hsu 18
While psychological symptoms often abate following natural disaster,Reference Felix and Afifi 6 , Reference Mills, Edmondson and Park 45 the ultimate improvement of post-disaster psychological sequelae has been linked to social and environmental factors experienced by survivors. Particularly, those exposed to unsafe neighborhoods, high crime rates, poor housing, and social isolation experienced persistent emotional distress.Reference Harville, Xiong, Pridjian, Elkind-Hirsch and Buekens 42 , Reference Lowe and Rhodes 46 Women, particularly low-income and racial minorities, are more likely to experience catastrophic property loss during natural disaster, leaving them vulnerable to higher rates of long-term displacement and corresponding loss of community ties post-disaster.Reference DeSalvo, Hyre, Ompad, Menke, Tynes and Muntner 47 This finding correlates directly with evidence that post-disaster psychological symptom improvement rates are slower and less consistent for women than for men.Reference Greenough, Lappi and Hsu 18 , Reference Harville, Xiong, Pridjian, Elkind-Hirsch and Buekens 42 Additionally, racial minorities in particular are likely to encounter greater difficulties recovering from natural disasters.Reference Jones, Ribbe, Cunningham and Weddle 44
Much less attention has been given to women’s experience of post-disaster IPV. Rates of IPV increase, often dramatically, after disaster.Reference Anastario, Larrance and Lawry 8 , Reference Larrance, Anastario and Lawry 11 , Reference Sastry and Gregory 48 For example, population estimates of GBV following Hurricane Katrina were as high as 17% for sexual abuse and 25% for IPV.Reference Anastario, Larrance and Lawry 8 Other studies, however, show contrary results, specifically where IPV following natural disaster has not increased.Reference Schumacher, Coffey, Norris, Tracy, Clements and Galea 49 , Reference Fagen, Sorensen and Anderson 50 In fact, some have found that natural disasters can in fact enhance a relationship.Reference Anastario, Shehab and Lawry 9 These studies repeatedly cite status as an internally displaced person,Reference Anastario, Larrance and Lawry 8 , Reference Fagen, Sorensen and Anderson 50 - Reference Jenkins and Phillips 52 loss of social networks and cultural integration,Reference Anastario, Larrance and Lawry 8 , Reference Frasier, Belton and Hooten 51 , Reference Clemens, Hietala, Rytter, Schmidt and Reese 53 and poverty/low socioeconomic status as aggravating factors. Protective factors against post-disaster IPV include non-urban settings, a strong sense of community and social cohesion among the studied population,Reference Schumacher, Coffey, Norris, Tracy, Clements and Galea 49 , Reference Frasier, Belton and Hooten 51 , Reference Jenkins and Phillips 52 and integration into the workforce.Reference Schumacher, Coffey, Norris, Tracy, Clements and Galea 49 , Reference Fagen, Sorensen and Anderson 50
These factors are often the same as those identified for women who experience post-disaster psychological trauma generally. The experience of post-disaster IPV and post-disaster negative mental health outcomes are not distinct: women who experience post-disaster IPV are significantly more likely to exhibit PTSD symptoms.Reference Anastario, Shehab and Lawry 9 In an integrative data analysis of disaster-related stressors experienced by survivors of Hurricane Katrina,Reference Cepeda, Valdez, Kaplan and Hill 3 the threat to physical integrity of self and others was identified as having the strongest association with posttraumatic stress and general psychological distress. Experiences of GBV increased from 4.6 to 16.3/100,000 per day in a displaced population following Hurricane Katrina. Reference Norris, Friedman, Watson, Byrne, Diaz and Kaniasty 7 This study also revealed that many of the women who are most at-risk for GBV and IPV also exhibited strong symptoms of depression, including sleep and appetite dysregulation, poor self-esteem, and suicidal ideation.Reference Norris, Friedman, Watson, Byrne, Diaz and Kaniasty 7 The 1993 Midwestern floods study revealed that, while positive intimate relationships may buffer individuals from the development of PTSD symptoms post-disaster, poor relationship adjustment is significantly associated with the development of PTSD symptoms.Reference Anastario, Shehab and Lawry 9
It should therefore come as no surprise that women who experienced post-disaster IPV and GBV particularly were vulnerable to major depressive disorder, increased symptoms of depression, and suicidal ideation.Reference Anastario, Larrance and Lawry 8 As an example, female survivors of Hurricane Katrina living in Federal Emergency Management Agency (FEMA; Washington, DC USA) travel trailer parks one year after the hurricane who reported post-disaster IPV were 10.4 times more likely to report symptoms of major depressive disorder than women without post-disaster IPV.Reference Anastario, Larrance and Lawry 8 Similarly, IPV post-disaster physical and psychological victimization were associated with hurricane-related PTSD.Reference Corrarino 13 In particular, certain PTSD symptoms are positively correlated with post-disaster experience of violence. For women living in FEMA travel trailer parks eight months after Hurricane Katrina, the odds of exposure to GBV generally were 3.8 times higher among women with appetite dysregulation, 2.3 times higher among women with low self-esteem, and 2.7 times higher among women with suicidal ideation in comparison to those without.Reference Norris, Friedman, Watson, Byrne, Diaz and Kaniasty 7
However, very little headway has been made into analyzing the long-term effects of post-disaster IPV on women’s health, particularly whether IPV ultimately improves for women following natural disasters. Some evidence shows that rates of IPV increase in the protracted time period post-disaster, especially if other exacerbating factors, such as displacement, are not relieved. In a study of survivors living in FEMA travel trailers following Hurricane Katrina, the overall prevalence of recent IPV increased from 2.5% six months after the hurricane to 7.6% two years after the hurricane.Reference Anastario, Larrance and Lawry 8 Similarly, a study of women exposed to massive flooding in the St. Louis, Missouri (USA) area found that, in the short-term, experience of loss resulting from natural disaster was correlated with positive relationship adjustment, while women who experienced negative relationship adjustment were more likely to exhibit PTSD symptoms.Reference Anastario, Shehab and Lawry 9 In general, however, the dearth of research in this field and the lack of generalizability of study results significantly impair the ability of policymakers to anticipate the needs of women who may still be experiencing increased rates of IPV several years after a natural disaster.
There may be hope. While women may experience stronger psychological effects than men, they may be more resilient, particularly when their social support networks remain accessible.Reference Wolf 33 In a study of evacuees who experienced the three consecutive wildfires in California (USA), Felix et al. found that, while women retrospectively reported greater disaster-related stressors and worse mental health at the time of disaster than their male counterparts, women’s social support networks played a direct and significant role in maintaining mental health; social support was not correlated to mental health at the time of fire or survey for males.Reference Davis, Grills-Taquechel and Ollendick 5 , 54 For pregnant and post-partum women who survived flooding of the Mississippi River in Iowa (USA) in 2008, both general social support and intimate partner support significantly impacted maternal mental health, mitigating the likelihood of perinatal depression.Reference Galea, Tracy, Norris and Coffey 55 , Reference Brock, O’Hara and Hart 56
Limitations
This scoping review was not without limitations. It is possible that some relevant articles were missed, despite the use of stringent and systematic search methodology as well as iterative steps to minimize any such omissions. Additionally, the inclusion and exclusion criteria employed, while carefully selected, may have been too narrow to include all relevant studies. Finally, while this review focused specifically on natural disasters in the US, findings are not generalizable to either man-made or technological disasters or disasters occurring outside of the US.
Conclusion
The best strategy to address the concerns highlighted in this study is through preparedness. The health consequences that women are likely to experience following natural disaster provide a useful guide for policymakers working to develop effective programs for disaster preparedness and relief. The evidence shown in this review that women are more likely to experience PTSD symptoms and other indicators of poor mental health should inform disaster preparedness planning surrounding the importance of offering early treatment and interventions. In particular, studies generating high levels of evidence, such as intervention development surrounding mental health care tailored to women after natural disaster, are a key component in successful disaster recovery. Awareness campaigns targeted towards decreasing IPV potentially can help decrease risk, as could including the need for women’s refuges in post-disaster planning. Finally, an intersectionality approach, where considering the cumulative impacts of race, gender, and economics, can help to guide effective interventions.
Supplementary Material
To view supplementary material for this article, please visit http://dx.doi.org/10.1017/S1049023X16000911
